Ice can get extraordinarily cold, far below the 0 °C (32 °F) freezing point most people associate with it. Once water solidifies, nothing stops the resulting ice from continuing to lose heat. In principle, ice can cool all the way down to absolute zero, −273.15 °C (−459.67 °F), the lowest temperature physics allows. In practice, ice at temperatures well below −200 °C exists naturally throughout the solar system and beyond, and the journey from freezing point to near absolute zero transforms ice in ways that go well beyond simply “getting colder.”
Why Ice Does Not Stay at Zero
A common misconception is that ice “lives” at 0 °C. That temperature is simply the point at which water and ice coexist at normal atmospheric pressure. While a mixture of ice and liquid water will hover near 0 °C as long as both phases are present (because the energy being removed goes into converting liquid to solid rather than lowering the temperature), once all the water has frozen, the ice is free to keep cooling. Put an ice cube in a freezer set to −18 °C and it will eventually reach −18 °C throughout. Place that same ice in a lab cryostat and you can bring it to −196 °C with liquid nitrogen, or to −269 °C with liquid helium. The ice remains solid throughout; it just gets colder.
Your home freezer keeps ice at roughly −18 °C (0 °F), which is why frozen food stays preserved. Commercial blast freezers go down to about −40 °C. Dry ice, which is solid carbon dioxide rather than water ice, sits at −78.5 °C and is sometimes used to chill water ice even further in shipping containers. Liquid nitrogen boils at −196 °C and can cool ice to that neighborhood in minutes. Each step down the temperature scale is progressively harder to achieve, because extracting heat from something already very cold requires increasingly specialized equipment.
What Changes Inside Ice as It Gets Colder
Cooling ice is not just a matter of making the same stuff colder. The crystal structure of ice actually shifts depending on temperature and pressure, and researchers have catalogued at least fifteen distinct solid phases of ice. The familiar form you find in your glass is called ice Ih (the “h” stands for hexagonal, after its crystal geometry). Under different conditions of pressure and temperature, ice molecules rearrange into other crystalline patterns with names like ice II, ice III, ice IX, and so on. At extreme pressures, there is even a phase called ice X in which individual water molecules no longer exist as such; instead, the oxygen and hydrogen atoms form a completely different bonded network.
At low temperatures and moderate pressures, the transitions between phases involve subtle rearrangements of how the hydrogen atoms (protons) are oriented. In ordinary ice Ih, the protons are somewhat disordered, pointing in various directions even though the oxygen atoms sit in a neat hexagonal lattice. Cooling ice Ih below about 72 K (−201 °C) with the help of a catalyst can coax it into ice XI, a proton-ordered version where the hydrogens line up in a regular pattern, giving the crystal ferroelectric properties.
At higher pressures, even more exotic transitions occur. Researchers have mapped phase boundaries between ice IX and ice XIII, finding that the transition pressure and temperature are interrelated. At around 0.30 GPa and 154 K (−119 °C), ice IX gives way to ice XIII, and the transition temperature drops further as pressure increases.
Negative Thermal Expansion
Most materials shrink when they cool, but ice bucks this trend in a subtle way at very low temperatures. Below about 60 K (−213 °C), the thermal expansion of ordinary ice actually becomes negative, meaning ice slightly expands as it cools further. This is an unusual property shared by only a handful of solids, and it arises from the way the vibrational modes of the crystal lattice shift at those temperatures. Computational models reproduce this behavior: the TIP4P/2005 water model, one of the more accurate molecular simulations of water, shows negative thermal expansivity below 60 K, matching experimental observations.
The effect is tiny in absolute terms and irrelevant for everyday life, but it matters to physicists trying to understand the fundamental forces inside ice crystals. It also means that characterizing ice as a simple, unchanging solid is misleading. Even well below the point where you would think “nothing interesting is happening anymore,” the internal structure of ice continues to evolve.
The Coldest Natural Ice in the Solar System
You do not have to build a laboratory to find extremely cold ice. The universe is full of it. On the Moon, permanently shadowed craters near the poles never see sunlight, and surface temperatures in these regions drop below 100 K (−173 °C). Water ice deposited there by comet impacts or solar wind chemistry can persist indefinitely at these temperatures. Subsurface ice buried about a meter deep beneath the lunar regolith remains stable up to roughly 131 K (−142 °C), because the overlying soil acts as insulation against temperature swings.
Farther out in the solar system, temperatures plunge much lower. On Pluto, surface temperatures hover around 40–50 K (−230 °C), and the presence of ammonia mixed with water ice lowers the freezing behavior even further. The eutectic temperature of the ammonia-water system is 176 K (−97 °C), meaning ammonia keeps some liquid persisting at temperatures nearly 100 K below the normal freezing point of pure water. This has implications for cryovolcanism on Pluto, where pressurized ammonia-water mixtures may occasionally erupt onto the surface.
The coldest natural ice is found on tiny dust grains drifting through interstellar space, far from any star. These grains carry thin coats of water ice at temperatures around 10 K (−263 °C), just ten degrees above absolute zero. At these temperatures, chemical reactions still occur on the icy surfaces. Radical species with low reaction barriers can combine even at 10 K, which is how complex organic molecules form in deep space, seeding the raw materials for future planetary systems.
Amorphous Ice and Glassy Water
When water freezes slowly under normal conditions, the molecules have time to arrange themselves into orderly crystals. But if you cool water fast enough, or deposit water vapor onto a surface cold enough, the molecules get trapped in a disordered jumble before they can organize. The result is amorphous ice, sometimes called glassy water, because it has the molecular randomness of a liquid locked into a solid form.
Amorphous ice is probably the most common form of water ice in the universe, since water vapor condensing onto frigid dust grains in space does not have the luxury of slow crystallization. Laboratory experiments that inject water vapor into cryogenic environments confirm that both crystalline and amorphous ice phases form depending on conditions like temperature and deposition rate.
Back in the 1970s, researchers discovered that many properties of supercooled water (liquid water chilled below 0 °C without freezing) show dramatic temperature-dependent behavior that appears to diverge near the homogeneous nucleation temperature, around −38 °C. This is roughly the coldest you can keep bulk liquid water before it spontaneously crystallizes. Below that threshold, if you want water to remain non-crystalline, you have to move into the amorphous ice regime, achieved by extremely rapid cooling (on the order of millions of degrees per second for small samples) or by vapor deposition at cryogenic temperatures.
Amorphous ice itself comes in several varieties. Low-density amorphous ice forms when vapor deposits below about 130 K. High-density amorphous ice can be made by compressing ordinary ice at low temperatures. A medium-density form was identified more recently. The relationships between these amorphous forms and the anomalous properties of supercooled liquid water remain an active and somewhat contentious area of research.
Quantum Weirdness Near Absolute Zero
As ice approaches absolute zero, quantum mechanical effects that are negligible at everyday temperatures start to dominate. One of the most striking is collective proton tunneling. In ordinary ice Ih, the hydrogen atoms (protons) sit along the bonds between oxygen atoms, and at higher temperatures they occasionally hop from one position to another, creating charge defects in the crystal. At very low temperatures, this classical hopping slows to a halt. But quantum mechanics offers another route: groups of six protons arranged in ring-shaped patterns within the hexagonal lattice can tunnel collectively, moving in concert as a delocalized quantum entity. This collective tunneling avoids creating the high-energy defects that single-proton hopping would produce.
Experimental evidence for this quantum behavior showed up in a surprising measurement. Researchers studying ordinary ice (H₂O) near 20 K (−253 °C) found a strange anomaly in its dielectric properties. When they repeated the experiment with heavy ice (D₂O, made with deuterium instead of hydrogen), the anomaly vanished. Because deuterium is twice as heavy as hydrogen, it tunnels far less readily. The disappearance of the anomaly in heavy ice pointed strongly to concerted proton tunneling as the cause, making ice one of the rare systems in nature where macroscopic quantum phenomena can be observed.
Even at absolute zero, ice retains a peculiar thermodynamic fingerprint. The third law of thermodynamics says that a perfectly ordered crystal should have zero entropy at absolute zero. Ice Ih violates this expectation: because the hydrogen atoms can sit in multiple orientations consistent with the ice rules (each oxygen has two hydrogens close and two far away), the crystal retains what is called residual entropy even at 0 K. Calculations and experiments agree on a value of about 0.81 calories per degree per mole, a number first estimated by Linus Pauling in the 1930s and refined in later computational work.
Why Extremely Cold Ice Matters in Technology
The ability to make and control very cold ice has practical consequences beyond pure science. One of the most important is in cryo-electron microscopy (cryo-EM), a technique that has revolutionized structural biology. In cryo-EM, a thin film of a biological sample in water is flash-frozen so rapidly that the water vitrifies, forming amorphous ice rather than crystals. Ice crystals would destroy the delicate structures of proteins and other biomolecules; amorphous ice preserves them in a near-native state.
The temperature at which samples are frozen and the cooling rate both affect whether the ice ends up crystalline or amorphous. Experiments with protein samples such as apoferritin have shown that thin aqueous films vitrify reliably in liquid ethane at calibrated temperatures below −110 °C, while most samples become crystalline above that threshold. Samples frozen at different temperatures within the vitreous range (−183 °C, −150 °C, and −110 °C) all yielded high-resolution protein reconstructions, with the −110 °C sample actually producing the sharpest images at 1.9 ångströms resolution.
Cryopreservation of biological tissues, from sperm and egg cells to donated organs, also depends on managing ice temperature and crystal formation. The goal in most cases is either to vitrify the sample entirely or to control where and how ice crystals grow so they do not puncture cell membranes. The science of how cold ice gets, and what form it takes at each temperature, underpins an entire industry of biobanking and fertility medicine.
How Close to Absolute Zero Can Ice Actually Get
Absolute zero is a theoretical limit, not a destination anyone has reached with any substance. The third law of thermodynamics implies that reaching exactly 0 K would require an infinite number of cooling steps. But researchers have come remarkably close. Laboratory cooling techniques can bring solid materials to within fractions of a millikelvin of absolute zero, and there is no reason ice could not be cooled to similar extremes given enough effort. At those temperatures, the behavior of ice is dominated entirely by quantum mechanics: zero-point vibrations persist even when all thermal energy has been extracted, because the uncertainty principle forbids the atoms from sitting perfectly still.
Early studies of thermal motion in ice and heavy ice confirmed that even at the lowest measured temperatures, there is residual atomic movement, a quantum zero-point energy that cannot be removed no matter how thoroughly you cool the sample. This is not a failure of the cooling process; it is a fundamental feature of matter at the quantum scale.
In interstellar space, the cosmic microwave background radiation sets a floor of about 2.7 K (−270.45 °C) for any object left long enough to reach thermal equilibrium with its surroundings. Ice on a dust grain in a particularly shielded molecular cloud might approach this limit, but it will not drop below it without active cooling. So the coldest natural ice in the universe sits at roughly 3 K, while the coldest laboratory ice can in principle be pushed far lower.
Ice on Worlds With Exotic Chemistry
When planetary scientists talk about “ice,” they do not always mean frozen water. On the outer planets and their moons, “ices” also include frozen methane, ammonia, nitrogen, and carbon dioxide. Each has its own freezing point and its own set of solid phases. But even focusing strictly on water ice, the conditions on other worlds produce forms of ice rarely or never seen on Earth’s surface.
On Pluto, the ammonia-water chemistry allows partial melting and refreezing at temperatures where pure water ice would be rock-solid. Researchers modeling Pluto’s interior have found that even small concentrations of ammonia (1–33% by mole fraction) can produce enough liquid during freezing to generate overpressures of 1–10 megapascals in a trapped crustal chamber, potentially driving cryovolcanic eruptions that reshape the surface.
Jupiter’s moon Europa and Saturn’s moon Enceladus have subsurface oceans beneath thick ice shells. The base of Europa’s ice shell, where it contacts the ocean, sits near the pressure-melting point, while the surface temperature is around 100 K. That means Europa’s ice shell spans a temperature range of roughly 170 degrees from bottom to top, and the ice at different depths may exist in different structural phases depending on local pressure and temperature. Understanding these phases is critical for planning future missions that aim to drill through the ice to sample the ocean below.
Residual Entropy and the Puzzle of Perfect Order
The residual entropy of ice Ih is more than a thermodynamic curiosity. It represents a genuine conflict between what ice does and what a simple reading of the third law of thermodynamics says it should do. In a perfectly ordered crystal at absolute zero, there should be only one way to arrange the atoms, giving zero entropy. But the ice rules allow an enormous number of nearly equivalent hydrogen arrangements, and ice Ih never fully sorts them out on its own, at least not on any timescale anyone has observed without adding a catalyst.
The proton-ordered phase ice XI represents what ice Ih “wants” to become at very low temperatures, and theoretical calculations support the idea that a proton-ordered structure is thermodynamically favored at sufficiently low temperatures. But the kinetics are agonizingly slow. Without a catalyst such as potassium hydroxide doping, the protons in ice Ih are essentially frozen in their disordered positions long before the crystal reaches the temperature at which ordering would be energetically preferred. This is a case where thermodynamics says “yes” and kinetics says “not in your lifetime.”
The interplay between residual entropy, proton ordering, and quantum tunneling at low temperatures makes ice one of the more philosophically interesting solids in condensed matter physics. A substance most people think of as simple and well-understood turns out to harbor deep puzzles that connect to fundamental questions about order, disorder, and the quantum behavior of matter at its coldest.